Unless the watermark is a giant “This is fake” diagonally across the center of the image, everything will get shared virally in the 3 second attention span of doomscrolling without a second look.
We already have open source image editing models, so how is this Google fault? This is UGC..
If I type something nefarious in Google search bar and then share that screenshot, is Google to blame for not even refusing rendering that text inside the input field?
That's a separate issue - of course it's only so that someone at google can tick a box "I've shipped an AI feature" to get promoted.
But we didn't gain any new ability to produce fake images, fake images can already be produced easily, even those that pretend they're from Google Earth. This was already possible, so I don't understand the concern here..
Are tech companies now responsible for everything the user does? That would end with total enslavement, don't go in that direction.
It could be used for "good things" too, as you could see in their promo post.
If I make a text editor, am I now responsible if someone uses it to write Mein Kampf? Back then people didn't complain that a notepad doesn't censor user input, but if we continue with your logic, we'll be surveiled constantly, I don't want that dystopian future to become reality.
instead of making potentially weak comparisons, could we consider why this feature was implemented in the first place? image editing models exist for good reason - I struggle to see a clear “why” for this product other than to cram generative content in yet another existing product.
If I type something nefarious in Google search bar and then share that screenshot, is Google to blame for not even refusing rendering that text inside the input field?